Messages in this thread | | | Date | Tue, 12 May 2020 23:00:59 +0200 | From | Thomas Gleixner <> | Subject | [patch V5 00/38] x86/entry: Entry/exception code rework - the leftovers |
| |
Folks!
This is V5 of the rework series. V4 can be found here:
https://lore.kernel.org/r/20200505135341.730586321@linutronix.de
Some of the non-x86 specific parts from part 1- 4 have been applied to the relevant tip branches and I merged them all together to base the rest on.
I have applied most of the patches from part 1-4 on top after fixing up the small review comments all over the place. The result can be found here:
git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git entry-base-v5
I dropped:
[patch V4 part 1 33/36] x86,tracing: Robustify ftrace_nmi_enter() [patch V4 part 2 10/18] x86/entry/64: Check IF in __preempt_enable_notrace() [patch V4 part 2 11/18] x86/entry/64: Mark ___preempt_schedule_notrace()
from V4 because Peter's attempt to sanitize ftrace_nmi_enter/exit() and the underlying callchains turned out to be incomplete and cause the whole __preempt_schedule_notrace() hackery which is caused by a gazillion of preempt_disable_notrace() / preempt_enable_notrace() invocations in the full call chain. I found a better way to do that and that's in part 5 now.
The remaining patches are mostly from part 5 of version 4 dealing with the #PF, device interrupts and the system vectors.
The main differences to V4 part 5 are:
- Provide nmi_enter/exit_notrace() which do not contain the ftrace_nmi_enter/exit() calls, use them in x86 and make the ftrace invocation explicit at a safe place.
- A fix for the reworked SVM code
- An optimization of asm_native_load_gs_index() which I found while staring at stuff in more detail
- The removal of the x86 specific irq_regs implementation which was aside of two underscores in the variable name exactly the same thing as the generic variant.
- Overhaul of the IRQ stack switching mechanics to avoid the objtool and unwinder nastiness of doing it in C inline ASM. As discussed here:
https://lore.kernel.org/r/20200507161020.783541450@infradead.org
We agreed on having the actual stack switch still in ASM and utilize it for device interrupts, soft interrupts and system vectors.
This comes for the price of an indirect call, but the device interrupt handling has an indirect call and soft interrupts and most system vectors are indirect call laden already.
There is a lightweight variant for the system vectors which do basically nothing, i.e. reschedule IPI and the KVM posted interrupt vectors. That variant runs on the interrupted stack and avoids the full entry/exit handling overhead by utilizing the conditional RCU variant of idtentry_enter/exit().
- Addressing a few review comments.
The series is also available from git:
git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git entry-v5-the-rest
Running the new objtool checks results in two warning types:
- The MCE code has an issue in the non-instrumentable section which is still under investigation and discussion
- When paravirt is enabled then there are a few complaints about invoking indirect calls (PVOPS) which leave the noinstr sections but they are halfways straight forward to fix.
Thanks,
tglx
8<--------------
arch/x86/include/asm/acrn.h | 11 arch/x86/include/asm/entry_arch.h | 56 -- arch/x86/include/asm/irq_regs.h | 32 - b/arch/x86/entry/calling.h | 25 - b/arch/x86/entry/common.c | 216 +++++++++- b/arch/x86/entry/entry_32.S | 260 +----------- b/arch/x86/entry/entry_64.S | 581 +++-------------------------- b/arch/x86/entry/thunk_64.S | 9 b/arch/x86/hyperv/hv_init.c | 9 b/arch/x86/include/asm/apic.h | 33 - b/arch/x86/include/asm/hw_irq.h | 22 - b/arch/x86/include/asm/idtentry.h | 262 ++++++++++++- b/arch/x86/include/asm/irq.h | 6 b/arch/x86/include/asm/irq_stack.h | 40 + b/arch/x86/include/asm/irq_work.h | 1 b/arch/x86/include/asm/irqflags.h | 10 b/arch/x86/include/asm/mshyperv.h | 13 b/arch/x86/include/asm/trace/common.h | 4 b/arch/x86/include/asm/trace/irq_vectors.h | 17 b/arch/x86/include/asm/traps.h | 21 - b/arch/x86/include/asm/uv/uv_bau.h | 8 b/arch/x86/kernel/apic/apic.c | 39 - b/arch/x86/kernel/apic/msi.c | 3 b/arch/x86/kernel/apic/vector.c | 5 b/arch/x86/kernel/cpu/acrn.c | 9 b/arch/x86/kernel/cpu/mce/amd.c | 5 b/arch/x86/kernel/cpu/mce/core.c | 4 b/arch/x86/kernel/cpu/mce/therm_throt.c | 5 b/arch/x86/kernel/cpu/mce/threshold.c | 5 b/arch/x86/kernel/cpu/mshyperv.c | 22 - b/arch/x86/kernel/head_64.S | 7 b/arch/x86/kernel/idt.c | 42 +- b/arch/x86/kernel/irq.c | 58 -- b/arch/x86/kernel/irq_64.c | 17 b/arch/x86/kernel/irq_work.c | 6 b/arch/x86/kernel/kvm.c | 14 b/arch/x86/kernel/nmi.c | 9 b/arch/x86/kernel/smp.c | 37 - b/arch/x86/kernel/tracepoint.c | 17 b/arch/x86/kernel/traps.c | 10 b/arch/x86/kvm/svm/svm.c | 2 b/arch/x86/mm/fault.c | 69 ++- b/arch/x86/platform/uv/tlb_uv.c | 2 b/arch/x86/xen/enlighten_hvm.c | 12 b/arch/x86/xen/enlighten_pv.c | 2 b/arch/x86/xen/setup.c | 4 b/arch/x86/xen/smp_pv.c | 3 b/arch/x86/xen/xen-asm_32.S | 12 b/arch/x86/xen/xen-asm_64.S | 4 b/arch/x86/xen/xen-ops.h | 1 b/drivers/xen/events/events_base.c | 6 b/drivers/xen/preempt.c | 2 b/include/linux/hardirq.h | 51 ++ b/include/linux/rcutiny.h | 1 b/include/linux/rcutree.h | 1 b/include/xen/events.h | 7 b/include/xen/xen-ops.h | 7 b/kernel/rcu/tree.c | 5 b/kernel/softirq.c | 35 + 59 files changed, 906 insertions(+), 1270 deletions(-)
| |